It has been established that early introduction of aggressive therapies in CD leads to better outcomes, compared to the standard step-up approach (D'Haens et al., 2008; Colombel et al., 2010). In particular, patients treated with early combination therapy (infliximab in combination with Azathioprine; D'Haens et al., 2008) experience longer periods of steroid-free remission, are more likely to achieve mucosal healing and, ultimately, to avoid surgical resection (D'Haens et al., 2008; Colombel et al., 2010). However, this strategy is not suitable for all patients because a non-negligible proportion of patients would have achieved prolonged remission even with the conventional treatment approach (Jess et al., 2007), and thus their treatment with combination therapy would expose them to unnecessary side effects and toxicity. For this reason, the ability to predict prognosis would be a major step toward improving care for patients with CD, and IBD generally (IBD Research Priority, 2015; Gerich et al., 2014).
A number of different variables have been associated with prognosis in CD; namely clinical factors, serological markers, and genetic variants (Gerich et al., 2014; Billiet et al., 2014). However, the predictive power of these markers, used alone or in combination, has proven to be limited so far, and none are suitable for routine use in the clinic (Loly et al., 2008; Markowitz et al., 2011; Ananthakrishnan et al., 2014). As a consequence, developing a reliable prognostic test for IBD, including CD and UC, remains a priority, and is currently recognised as one of the most important unmet needs in gastroenterology (IBD Research Priority, 2015).
To accomplish this goal, genetic variants are promising candidate prognostic markers, because they are stable, can be easily measured, and because the genetic architecture of CD has already been studied extensively, at least with regard to susceptibility (Jostins et al., 2012). However, the genetic factors discovered so far may only explain 5-6% of the phenotypic variance observed in CD outcome between patients. Even if this figure were to increase, as better-powered studies discover other outcome-associated variants, it seems unlikely that genetic factors will be sufficient, in isolation, to enable accurate prediction of outcome in CD, and IBD generally. This is consistent with the notion that environmental factors, including smoking, play an important role in the natural history of IBD (Kaser et al., 2010).
Gene expression markers may be better candidates to overcome these limitations, particularly if measured directly in tissues or cell types which are involved in the disease pathogenesis (McKinney et al., 2010; Lee et al., 2011). In fact, gene expression may capture more information about interactions between organism and external environment (Choi et al., 2007), while still reflecting some aspects of individual genetic background. On the other hand, gene expression is not normally expected to be stable over time, and isolating particular cell populations without affecting gene expression levels is technically challenging (Lyons et al., 2007). Outside a controlled research environment, all of these factors limit the use of gene expression markers in a clinical context.
It has recently been reported that a distinctive gene expression signature, detectable in CD8+ T cells, can be used to predict prognosis in IBD (Lee et al., 2011). In particular, such a signature was able to identify, at the time of diagnosis and before therapy, two distinct groups of patients (IBD1 and IBD2), associated with different clinical courses in both CD and UC (Lee et al., 2011). More precisely, when managed with the standard step-up approach (Peyrin-Biroulet et al., 2008), patients classified as belonging to the IBD1 group consistently showed significantly higher risk of treatment escalation than patients in the IBD2 group (Lee et al., 2011), thus providing a rationale for treating these patients with more aggressive therapies at an earlier stage (Lee et al., 2011).
The aforementioned signature represented a major advance towards prediction of outcome in IBD (Friedman et al., 2011) for three main reasons. Firstly, the difference in outcome between patient strata is marked enough to be potentially useful to guide therapeutic decisions (Lee et al., 2011; Friedman et al., 2011), unlike previously reported prognostic markers (Gerich et al., 2014). Secondly, an overlapping CD8+ gene expression signature was previously reported to predict disease outcome in Systemic Lupus Erythematosus (SLE) and ANCA-associated vasculitis (AAV) patients, thus suggesting the hypothesis that common biological processes may underlie the prognosis of different inflammatory diseases (McKinney et al., 2010). Finally, a partial but compelling mechanistic explanation for the underlying biological processes was recently proposed (McKinney et al., 2015). These last two points are particularly important. In fact, investigating prognosis necessarily requires longitudinal studies where data collection is both time consuming and expensive, thus limiting the study size. As a consequence, it often remains in doubt whether a small cohort of patients can capture enough complexity from the underlying population and whether the proposed prognostic marker is indeed reproducible. In view of this, observing consistency with regard to disease outcomes across different cohorts of patients with different diseases, together with credible mechanistic insights, strongly increases the confidence in the reproducibility of the aforementioned CD8+ T cell gene expression signature (McKinney et al., 2010; Lee et al., 2011).
However, an important problem remains to be solved before the CD8+ T cell gene expression signature can be routinely used to stratify patients in a clinical setting. Assigning a patient to IBD1/IBD2 groups currently requires RNA extraction from purified CD8+ T cells (McKinney et al., 2010; Lee et al., 2011). It has been repeatedly observed that this step adds a considerable amount of complexity to a potential prognostic test, thus limiting its applicability to small numbers of samples in a controlled research setting (Friedman et al., 2011; Billiet et al., 2014).
On the contrary, being able to detect the IBD1/IBD2 subgroups in a readily accessible biological sample, such as whole blood, would greatly facilitate its clinical utility and, potentially, its applicability as a prognostic marker. Furthermore, because whole blood samples, but not purified CD8+ T cells, have been routinely collected during some clinical trials, this would open the possibility of reanalysing past IBD drug trials in order to re-evaluate drug efficacy after patient stratification.
Despite being potentially useful, it should be noted that detecting the IBD1/IBD2 signature in samples different from purified CD8+ T cells has previously proven to be challenging (Lee et al., 2011). In fact, it was repeatedly observed that gene expression signatures derived from CD4+ T cells could not be used to stratify patients by disease outcome in IBD (Lee et al., 2011), SLE, or AAV (McKinney et al., 2010). Consistent with these observations, it was not possible to identify equivalent prognostic signatures in peripheral blood mononuclear cells (PBMC) using the same unsupervised clustering methods (Monti et al., 2003) originally used to discover the CD8+ T cell expression signature (McKinney et al., 2010; Lee et al., 2011). This is may be due to the fact that CD8+ T cells represent a very small and variable fraction of the PBMC population (Lyons et al., 2007).
Moreover, while prognostic signatures can be discovered using high throughput gene expression profiling technologies, such as microarrays (Schena et al., 1995), a viable prognostic test needs to rely on a smaller scale gene expression platform such as real time quantitative PCR (RT-qPCR) (Freeman et al., 1999). In fact, most of modern prognostic tests developed for different conditions, such as AlloMap, Oncotype Dx and CorusCAD are qPCR based tests (Micheel et al., 2012), while only a few older tests, such as MammaPrint, rely on microarrays (Micheel et al., 2012). For this reason, if a signature that recapitulated the IBD1/IBD2 subgroups could be discovered in whole blood, the possibility of detecting it by qPCR would be crucial for its application in a clinical setting.
There thus remains a need in the art for an IBD1/IBD2 gene expression signature which can be detected in whole blood using methods such as qPCR.
Statements of Invention
The present inventors have surprisingly identified a gene expression signature that can be used to assess whether an individual is at high risk or low risk of IBD (e.g. ulcerative colitis or Crohn's disease) progression by analysing gene expression in a whole blood sample obtained from the individual. The expression of these genes in whole blood can be determined, for example, by RT-qPCR. Specifically, the present inventors have discovered that a high risk (IBD1) phenotype is characterised by upregulated expression of genes ARRDC4, GBPS, P2RY14, VTRNA1-1, IL18RAP, HP, NUDT7, GZMH, TRGC2, and GZMK, and downregulated expression of genes LGALSL, FCRL5, IF144L, LINC01136, LY96, and TRGV3 in whole blood, relative to the level of expression of these genes in an individual who has a low risk (IBD2) phenotype.
In one aspect, the present invention thus provides a method of assessing whether an individual is at high risk or low risk of IBD progression comprising establishing, by determining the expression level of two or more genes in a whole blood sample obtained from the individual, whether said individual has a high risk (IBD1) or low risk (IBD2) phenotype, wherein the two or more genes are selected from the group consisting of: ARRDC4, GBP5, P2RY14, VTRNA1-1, IL18RAP, HP, NUDT7, GZMH, TRGC2, LGALSL, FCRL5, IF144L, LINC01136, LY96, GZMK, and TRGV3.
In one embodiment, the method may comprise determining the expression level of said two or more genes using e.g. RT-qPCR, digital PCR, or whole transcriptome shotgun sequencing. In this case, the method may comprise: (i) providing a whole blood sample obtained from the individual; (ii) extracting RNA (e.g. mRNA) from the whole blood sample; (iii) converting the RNA (e.g. the mRNA) into cDNA, and (iv) performing RT-qPCR, digital PCR, or whole transcriptome shotgun sequencing to determine the expression level of the two or more genes.
The present invention also provides an autoimmune disease progression risk assessment system for use in a method of the invention, wherein the system comprises a tool or tools for determining the expression of two or more genes selected from the group consisting of: ARRDC4, GBP5, P2RY14, VTRNA1-1, IL18RAP, HP, NUDT7, GZMH, TRGC2, GZMK, LGALSL, FCRL5, IF144L, LINC01136, LY96, and TRGV3, and a computer programmed to compute an IBD progression risk score from the gene expression data of the subject.
The tool(s) for determining expression of the genes may be, or comprise, reagents for establishing the expression level of the genes in question using any technique described herein, such as RT-qPCR, digital PCR, microarray analysis, whole transcriptome shotgun sequencing, or direct multiplexed gene expression analysis. For example, the tool(s) may be, or comprise, primers suitable for establishing the level of expression of the genes in question using e.g. RT-qPCR, digital PCR, whole transcriptome shotgun sequencing, or direct multiplexed gene expression analysis. The design of suitable primers is routine and well within the capabilities of the skilled person. Where the method comprises multiplexed gene expression analysis, the tool(s) may in addition, or alternatively, include fluorescent probes for establishing the level of expression of the genes in question. The tool(s) may also be, or comprise, RNA extraction reagents and/or reagents for reverse transcription of RNA into cDNA. The tool(s) may also be, or comprise, one or more articles and/or reagents for performance of the method, such as buffer solutions, and/or means for obtaining the test sample itself, e.g. means for obtaining and/or isolating a sample and sample handling containers (such components generally being sterile). The computation of an IBD progression risk score may be achieved in a number of ways and exemplary methods are set out below.
The present invention further provided is a method of treating IBD in an individual. In one embodiment, the method may comprise: (i) identifying the individual as one who is at high risk or low risk of IBD progression using a method of the invention, and (ii) subjecting the individual to treatment for IBD.
In an alternative embodiment, the method may comprise: (i) requesting a test providing the results of an analysis to determine the expression level of two or more genes in a whole blood sample obtained from the individual, wherein the two or more genes are selected from the group consisting of: ARRDC4, GBP5, P2RY14, VTRNA1-1, IL18RAP, HP, NUDT7, GZMH, TRGC2, GZMK, LGALSL, FCRL5, IF144L, LINC01136, LY96, and TRGV3, and (ii) treating the individual determined to have an IBD1 or IBD2 phenotype for IBD.
Also provided is a kit for assessing whether an individual has a high risk IBD1 or low risk IBD2 phenotype, wherein said kit comprises reagents for establishing the expression level of two or more genes selected from the group consisting of: ARRDC4, GBP5, P2RY14, VTRNA1-1, IL18RAP, HP, NUDT7, GZMH, TRGC2, GZMK, LGALSL, FCRL5, IF144L, LINC01136, LY96, and TRGV3.
The present invention also relates to an in vitro method of identifying a substance capable of inducing a low risk (IBD2) phenotype in an IBD patient. The method preferably comprises: providing (i) a whole blood sample obtained from an IBD patient prior to treatment with a substance of interest, and (ii) a whole blood sample obtained from the IBD patient following treatment with the substance of interest, wherein the IBD patient has been determined to have a high risk (IBD1) phenotype using a method of the invention; and determining the expression level of two or more genes selected from the group consisting of ARRDC4, GBP5, P2RY14, VTRNA1-1, IL18RAP, HP, NUDT7, GZMH, TRGC2, GZMK, LGALSL, FCRL5, IF144L, LINC01136, LY96, and TRGV3 in samples (i) and (ii); wherein a lower expression of genes ARRDC4, GBP5, P2RY14, VTRNA1-1, IL18RAP, HP, NUDT7, GZMH, TRGC2, and GZMK, and a higher expression of genes LGALSL, FCRL5, IF144L, LINC01136, LY96, and TRGV3 in sample (ii) relative to sample (i) indicates that the substance is capable of inducing a low risk (IBD2) phenotype in an IBD patient.
The present invention further relates to an in vitro method of identifying a substance capable of treating IBD in an individual. The method preferably comprises: (i) identifying an individual who is at high risk or low risk of IBD progression using a method of the invention; and (ii) comparing the level of IBD progression in the individual following treatment with the substance of interest with a control, wherein a lower level of IBD progression in the individual compared with the control indicates that the substance is capable of treating IBD.